Social Capital and Community Preparation for UrbanFlooding in China

Alex Y Lo

The Kadoorie Institute, University of Hong Kong, Hong Kong, China

Bixia Xu

Griffith School of Environment, Griffith University, Australia

Faith Chan

School of Geographical Sciences, University of Nottingham,Ningbo, China

Ruixian Su

School of Urban and Environmental Studies, Tianjin NormalUniversity, China

Citation:

Lo, A.Y., Xu B., Chan F.K.S. and Su R. (2015) Social capital and community preparationfor urban flooding in China. Applied Geography 64: 1-11.

DOI: http://dx.doi.org/10.1016/j.apgeog.2015.08.003

1

Abstract

Social capital can enhance community resilience to environmentalchange. Productive and trusted relations among social actors andeffectual social norms can help local residents share resources,information and risks. The main objective of our study is tounderstand the ways in which social attributes and riskconsiderations influence adoption of resilient economic measuresby individuals for reducing potential losses due to catastrophicrainstorm and flooding. This article provides evidence from Chinaon how social capital contributes to anticipatory adaptation toenvironmental change. The inquiry is based on structuredinterviews with local residents of Tianjin, a flood-prone portcity in China, and a standard regression analysis. Findings showthat the intention to make preparation increases with the levelsof social expectation, social relationship, and institutionaltrust. Perceived risk and damage experience, however, have nosignificant impacts. This suggests that building social capacityand trust will be more effective in enhancing communityresilience than merely increasing awareness of hazard risks. Wecall for greater efforts on strengthening the capacity of formaland informal communal institutions. The structural changesrequired, however, are challenging.

Keywords: social capital; community resilience; climate change adaptation; risk mitigation; flooding; China

2

Introduction

Swathes of land are exposed to climate change disturbance as

average temperatures are rising in the coming decades. The

warming climate brings more energy into the meteorological system

and results in a more vigorous hydrological cycle (Güneralp, et

al., 2015). Some regions will receive more extreme precipitation

and torrential rains. Coastal low-lying built-up areas are likely

to experience rainstorm waterlogging and inundation more often

than before. Older capital cities, in which human activities and

settlements are concentrated and urban drainage systems have

reached their capacity, are particularly vulnerable.

Middle-income countries, such as China, are economic victims

of abrupt climate change. According to the World Bank (2010),

natural hazards in the past 40 years accounted for a total of

US$2,300 billion in economic costs (in 2008 dollars). Middle-

income countries incur the greatest proportional damage – more

than poorer countries with few assets and richer countries which,

with more capital, can more effectively prevent damage. Beijing,

the capital of People’s Republic of China, recorded a total

economic damage of over ¥10 billion Chinese Yuan (approx. USD$1.5

billion) and a large number of casualties (77 deaths and 1.6

million people affected) in a catastrophic rainstorm event in

2012 (Wu, 2012). Other megacities in China, such as Shanghai,

Guangzhou, Tianjin, and Shenzhen, are also exposed to high risks

of rainstorm waterlogging, flooding and/or hurricanes (Xie, et

3

al., 2014). The changing climate threatens to increase the

economic burdens on residents and communities by an order of

magnitude.

Enhancing resilience to these impacts is a pressing issue.

This can be achieved by encouraging residents to adjust household

plans in ways that can reduce their potential economic losses

from future extreme weather events, such as taking out flood

insurance and diversifying income sources (Goulden, et al.,

2013; Keil, et al., 2008; Lo, 2013). In developing countries,

however, individuals and households encounter many barriers to

adapting to climate change (Le Dang, et al., 2014; Suckall, et

al., 2014). One of these barriers is the lack of, or

misperception of, information on hydrological risks. It is often

assumed that the intention to make preventive arrangements for

flooding is a function of perceived risk (Kunreuther, 1996,

2006). Misperception and myopic attitudes would result in delays

or failures in undertaking preventive actions ahead of time to

reduce the potential impacts of natural hazards on households and

communities. In reality, many ordinary people tend to see the

likelihood or severity of a natural hazard event causing damage

to their property and affecting their livelihood as being

sufficiently low. Systematic misperception of flood risks is

identified as the main reason for poor or inadequate preparation

for catastrophic flooding events (Kunreuther, 1996, 2006; Miceli,

et al., 2008).

4

However, recent research has suggested that the importance

of risk perception has been overstated. Based on a German case

study, Bubeck et al. (2013, p. 1336) conclude that “risk

perceptions rarely relate significantly to improved flood

mitigation behaviour”. An Australian survey reported by Lo (2013)

has come to the same conclusion. More evidence can be found in

Bubeck et al.’s (2012) meta-analysis of 16 relevant empirical

studies. Rather than taking risk perception (and hazard

experience) for granted, researchers have increasingly turned

their focus toward social capital, which refers to the structure

of relations among social actors (Coleman, 1990; Pretty & Ward,

2001). Social capital can help enhance the capacity of households

and communities to cope with climate change impacts (Bihari &

Ryan, 2012; Goulden, et al., 2013), although in some cases it

contributes to vulnerability (Wolf et al., 2010). Variations in

the level and form of social capital could explain people’s

willingness to make hazard adjustments. The quality of social

norms and interaction with people around and the community has

significant implications for anticipatory adaptation to climate

change (Eriksen & Selboe, 2012; Frank, et al., 2011).

This research is situated in the context of a developing

country in search of more effective ways for enhancing community

resilience to extreme weather events and reducing burdens on

local communities. We focus on voluntary economic preparation at

the household level, which is an area of research that requires

more attention than it has received, given its practical

5

implications for state economic planners and aid organizations

(Alinovi, et al., 2009; World Bank, 2010). Local research efforts

remain inadequate in China, which is home to a great number of

flood-prone and densely populated cities and towns. Existing

Chinese studies have centred on the effects of human cognition

and hazard experience in driving behavioral response (Ge, et al.,

2011; Huang, et al., 2010; Lo & Cheung, 2015; Yu, et al., 2013;

Zheng & Dallimer, 2015). Few attempts have been made to utilize

the concept of social capital. We bring the idea of economic

resilience (Alinovi, et al., 2009; Rose, 2004, 2007) together

with investigations of the role of social capital in fostering

behavioral adjustments to environmental change. Findings will be

useful for identifying critical factors influencing the intention

to act and formulating strategies for promoting risk management

and preparation for extreme weather events at the household

level.

This paper reports a quantitative study aimed at examining

how social attributes and risk considerations influence voluntary

adoption of resilient economic measures by individuals. Evidence

was solicited from structured interviews with local residents of

Tianjin, China. In the next section, we further elaborate on the

conditions and utility of social capital. Then, we introduce the

study area and survey instrument, followed by a statistical

analysis of survey data. In the last section, we summarize

findings and reflect on the issues being addressed.

6

Social capital and adaptation to environmental change

Social capital captures the idea that social relations and social

norms form the basis of sustainable communities. It is generated

and accumulated in the processes of social interaction and

engagement with social networks and institutions, both formal and

informal ones, yielding a wide range of economic and societal

benefits. While Coleman (1990) understands social capital as a

largely unintentional outcome of social interactions and

organization, Putnam (1995, p. 664-665) defines it as an enabler

of collective action: “features of social life – networks, norms

and trust – that enable participants to act together more

effectively to pursue shared objectives”. Pelling and High (2005)

describe Putnam’s (1995) definition as the most frequently used

and widely accepted one, which is therefore adopted for the

present paper.

Social scientists have found the concept with multiple,

analytically distinctive dimensions. Pelling and High (2005, p.

310), for example, note that social capital consists of two

complementary components, namely, interpersonal relationship, and

trust and reciprocity. Interpersonal relationships are developed

from social ties, networks, and connections. Extensive and close

interpersonal relationships are an outcome of trust and

reciprocity, which reduce transaction costs and encourage

cooperative behavior. Conformity to common expectations may also

7

nurture social capital. As Pretty and Ward (2001, p. 211)

suggest, the four critical aspects of social capital are 1)

relations of trust; 2) reciprocity and exchanges; 3) common

rules, norms and sanctions, and; connectedness, networks and

groups. Adger (2003, p. 389) describes the concept in similar

terms, encompassing relations of trust, reciprocity and exchange,

evolution of common rules, and networks. The first, second and

fourth dimensions outlined by Pretty and Ward (2001) and Adger

(2003) match the two conceptual components identified by Pelling

and High (2005), while the third dimension has nuanced

connotations. Common rules, norms and sanctions render other

people’s action predictable, reduce uncertainties in the outcomes

of a particular action, and give individuals the confidence to

invest in related activities. These manifestations of social

capital are conducive to motivating collective action that would

otherwise be deemed to be too costly to undertake.

Social capital among members of a community is a key

determinant of its vulnerability and resilience to environmental

changes and uncertainties (Adger, 2000, 2003; Barnett & Eakin,

forthcoming; Frank, et al., 2011; Goulden, et al., 2013; Pelling,

2011; Wilson, 2015). Adger (2003, p. 401) has argued that “many

aspects of adaptive capacity reside in the networks and social

capital of the groups that are likely to be affected”. The

structure and quality of social relations select what impending

changes to act upon and determine the type and range of options

for coping with these changes. Frank et al. (2011) find that the

8

ways in which Mexican farmers understand and respond to climate-

related threats to their welling and livelihoods are shaped by

their perception of themselves in relation to others in the

community, whereas risk perception and experience alone do not

have significant influence on the motivation to adopt adaptive

measures. Similarly, Goulden et al. (2013) report that the social

connections among Ugandan villagers contribute to household

resilience by facilitating livelihood diversification through

providing access to resources and livelihood opportunities,

access to credit, reducing costs of engaging in activities,

facilitating migration, etc. Social capital links give households

more opportunities for mitigating and spreading the risks of

environmental change.

Yet, the accumulation of social capital is not necessarily a

social ‘good’ and may create perverse incentives undermining

adaptive capacity (Pelling and High, 2005; Pelling, 2011).

Established networks, norms and trust, in Putnam’s (1995) terms,

may perpetuate social relations and practices that are counter-

productive or detrimental to adaptation to condition changes. For

example, in Norway, the cultural and social norms of appreciating

local history and recalling the collective memory about the home

town in which local residents live had contributed to the social

organization of denial of climate change, which has a salient

orientation towards long-term futures and requires forward-

thinking (Norgaard, 2011). High levels of trust in the state,

especially post-communist regimes that used to assume full

9

responsibilities for central planning (Vari, et al., 2003), may

result in an expectation among potential victims that generous

post-disaster reliefs and home reconstruction assistance would be

provided if natural perils occur, and consequently, lead to

delays in undertaking preventive actions, a common problem known

as ‘moral hazard’. Wolf et al.’s (2010) British case study shows

that the social networks of elderly people could perpetuate

narratives of independence, capability and resilience and

consequently contribute to their vulnerability to heat waves

rather ameliorating it. Thus, the structure of social relations

and practices has mixed impacts on subjective experience with

condition changes and voluntary adoption of adaptive measures.

The relationship between social capital and adaptation to climate

change is complex and not invariably a positive one (Pelling,

2011; Wolf, et al., 2010).

Therefore, the notion that social capital is instrumental to

fostering proper behavioral adjustments to the prospects for

extreme weather events cannot be taken as given. This issue

requires further investigation in the context of China, where the

state continues to dominate urban management and assume the

primary responsibility for managing flood risks, with the

possible consequence of displacing the motivation of the general

public to manage these risks. Our study sought to ascertain how

social capital attributes influence voluntary adoption of

resilient economic measures in a Chinese urban community. To

examine the claim that adoption of such measures is poorly

10

informed by risk perception and hazard experience (Bubeck, et

al., 2012; Bubeck, et al., 2013; Frank, et al., 2011; Lo, 2013),

we also included these elements in the analysis. The inquiry was

guided by two working hypotheses: Intended economic preparation

for catastrophic rainstorm waterlogging and flooding is

significantly related to 1) risk perception and hazard experience

and 2) attributes of social capital. A case study of Tianjin is

introduced below.

Urban flood hazards in Tianjin

Tianjin is located in the northeastern part of North China Plain

area and adjacent to Beijing and Hebei Province (Figure 1). The

city is the fourth largest in China by urban population and the

third largest by urban area, consisting of six urban districts,

along with four suburban districts and five counties. Tianjin has

experienced phenomenal socio-economic growth since China embarked

on the far-reaching economic reforms in the late 1970s, and

continues to be a major industrial powerhouse in the country. In

2013, the city recorded a high GDP growth rate at 12.5 per cent

(Tianjin Bureau of Statistics, 2014) and GDP per capita ¥97,622

Chinese Yuan (approx. USD $14,360) (National Bureau of Statistics

of China, 2014). A large number of job-seeking migrants from

rural China have moved to Tianjin, driving housing needs and

urban sprawl (Chan, et al., 2012). Municipal population has risen

from 9.3 million in 1993 to over 14.7 million in 2013 (Tianjin

11

Bureau of Statistics, 2014). Although urban development can

enhance people’s well-beings by driving infrastructural

development (e.g. mass transit), it may reinforce some

environmental threats, such as increasing the risks of rainstorm

waterlogging and flooding.

The city of Tianjin is situated at the outlet of a major

tributary, namely Yongding River in Haihe River basin. Fluvial

discharge passes through the city to Bohai all the way from

Beijing and the Northern China, putting it under threat of

fluvial (riverine) flooding. The urban drainage system is likely

to reach its capacity on the release of peak flow during the

storm period, because rapid urbanization has increased the area

of concrete surface and diminished penetration or infiltration of

the surface runoff into soil (Han, et al., 2006). Moreover,

massive urbanization and new coastal settlements have contributed

to dramatic land-use changes in Tianjin. Grey infrastructure has

replaced vast areas of blue-green sites that naturally restore

stormwater, such as forest and grassland, and regulate coastal

tides and surges, such as wetlands. The loss of wetland, in

particular, accelerates land subsidence in the city (Yi et al.,

2014). Land subsidence causes gradual shrinking of ground-surface

elevation, consequently increasing the risk of coastal flooding

and sea water intrusion during dry season (Wang, 1998). These

hazards are aggravated by the rising sea levels along the Bohai

and Yellow River areas by approximately 2.2mm per year (Varis, et

al., 2012). These factors combined, including increasing

12

population and economic activities, land-use changes,

accelerating sea level rises and surges, have increased the risks

of flooding confronted by Tianjin (Fuchs, et al., 2011; Wang, et

al., 2014).

On 21st July 2012, the heaviest rainstorm in six decades hit

Beijing and caused the tragic death of 77 residents (Wu, 2012).

Cloudbursts and floodwaters also passed through Tianjin, which is

114 km to the southeast of Beijing. The Tianjin city was ravaged

by torrential rain by days and experienced two hits of heavy

rainfalls. During the first heavy rain event from 23:30 on the

21st July 2012 until 6:00 of the next day, Tianjin city centre

recorded an average rainfall of 74 mm. Rainfall in Hebei and

Hongqiao, two of the six urban districts of Tianjin, exceeded

official ‘Heavy Rainstorm’ Levels. Acute and prolonged intensive

rainfall resulted in severe waterlogging in 18 places of the city

centre, causing significant damage on properties and

infrastructure (Tianjin Water Authority, 2012).

Tianjin suffered from a second heavy rainstorm hit in the 25th

and 26th of July 2012. The central and southern areas received

‘Rainstorm’ to ‘Heavy Rainstorm’ levels. There were 28 automatic

rainfall monitoring stations in Tianjin. Twenty six of these

stations reported accumulated rainfall of over 100 mm, 17

stations reported rainfall of over 150 mm with the peak value of

176.8 mm found at Tuanjiebeili Station during the 17-hour storm

that started on 25th July at 7:00 and lasted until mid-day of the

13

next day. An area of nearly 160 km2 suffered waterlogging and 69

places exceeded 300 mm of rainfall (Li, 2014, p.32-33). With an

average of almost 147.6 mm of rain falling in the city of 14.7

million people, the deluge breaks records for rainfall in Tianjin

since the historical flooding in 1963 and 1996 (Tianjin Water

Authority, 2008). Part of the Jixian County, a suburb in northern

Tianjin, was inundated, and 1106 residents and 260 tourists were

forced to evacuate (Li, 2014, p.33). The heavy rainstorm

disrupted telecommunication in Jixian area and wreaked havoc at

the Tianjin Binhai International Airport. More than 54 inbound

and outbound flights were rescheduled or cancelled (Li, 2014, p.

33). The ‘1-in-100-year’ rainstorms and urban inundation raise

concern about the capacity of Tianjin’s drainage networks for

handling massive deluge. These catastrophic events have stirred

up public criticisms against city managers and the quality of

urban infrastructure amid rapid urbanisation.

Climate change complicates flood risk management in Tianjin

as these urban flood hazards are likely to become more frequent

and intense in the coming years. Local communities and households

are exposed to escalating risks of property damage and rising

economic losses, due to the rapid socio-economic development in

the larger Beijing-Tianjin-Hebei region and increases in the

values of economic activities and assets. It is imperative to

promote voluntary adoption of resilient economic measures to

reduce the damage and recovery costs borne by city management and

14

the households themselves. Our study was designed to shed light

on this endeavor.

Research methods

Survey questions

The study involved a structured questionnaire survey of Tianjin

residents. The questionnaire collected four sets of information.

The first one was the respondents’ motivation to make preventive

arrangements for coping with disruptive urban flood hazards in

the Tianjin city. Respondents were asked to express their

intention to adopt strategies that could reduce potential

economic losses due to catastrophic rainstorms and flooding. Five

close-ended questions were included, each presenting an attribute

of economic resilience, namely, consumption elasticity (reducing

expenditures), access to insurance (recovering damage), asset

mobility (reducing losses), income diversity (maintaining

economic functioning), and asset distribution (reducing losses).

All of these questions were read in the context of urban flood

hazards and preceded by a brief discussion with the interviewer

on the past and future rainstorm or flooding events in Tianjin.

The design of the question statements was informed by key

research reports, notably Alinovi et al. (2009) and World Bank

(2010). Responses were measured on a five-point Likert scale.

Higher scores denote stronger intention to adopt resilient

15

economic strategies as a means to reduce potential losses in the

event of catastrophic rainstorms and flooding.

Another set of survey items recorded respondents’ experience

with rainstorm and flooding events and their perception of these

risks. A binary (yes/no) item ascertained whether the respondents

had incurred any damage from a rainstorm or flooding event,

followed by a five-option qualitative question inquiring the

extent of damage. The ensuing questions were used to measure

perceived likelihood and severity of severe urban inundation in

the future: how likely is Tianjin to experience a major

catastrophic event like the 2012 rainstorms? and how much

personal damage would it cause? Respondents were then requested

to consider how likely such hazard events would become more

frequent and intense in the next two decades. All items, except

the first one, were measured on a five-point scale.

Social capital variables were constructed broadly following

Putnam’s (1995) definition, encompassing aspects of networks,

norms and trust in social life. Eight specific survey items

gauged the strength of social-normative effects on respondents

(e.g. “do you think your family would encourage you to prepare

for urban flood hazards?”), the quality of social relationships

(e.g. “how is your relationship with your friends?”), and the

level of trust in institutions (e.g. “do you trust that your

local residential committee can help residents cope with the

impacts of urban flood hazards?”). Responses were recorded on a

16

five-point scale. Exact wording and response options can be found

in Appendix. Since these questions solicited evidence on social

capital from a subjective point of view (i.e. the respondents’

self-assessment), care is required when comparing with other

studies that use objective, behavioral indicators, such as

membership in social organizations.

Personal characteristics were included in the statistical

analysis as control variables. Four questions recorded

respondents’ age, education, sex, and monthly income. An

additional one asked whether their current home is located at the

ground floor of a residential building (the great majority of

residents in Tianjin’s core city area live in densely packed

multi-storey buildings). This is an important control measure

because ground-floor units and assets are vulnerable to

inundation and their occupants may be predisposed to risk

mitigation.

Data collection

The study sites are located at the six central administrative

districts of the Tianjin city, namely, Heping, Hedong, Hexi,

Nankai, Hebei, and Hongqiang, which are exposed to threat of

flooding and have a very high concentration of people,

infrastructure and economic activities. Figures 2 and 3 show a

selected view of the study sites, focusing on the urban form and

the ground setting respectively.

17

[FIGURES 2 and 3 ABOUT HERE]

Face-to-face interviews were conducted in the public areas.

Sample was distributed amongst these districts approximately in

proportion to local population size. Two or three major public

parks or neighbourhood parks were identified from each of the

selected districts as study sites. The selection was based on

proximity to housing clusters, pedestrian traffic, and

feasibility for conducting intercept survey. Park visitors were

invited to complete the questionnaire, which lasted about 8 – 10

minutes each. A certain number of passers-by were also

inadvertently included, since urban parks in Tianjin’s core city

area are tightly integrated with the street footpath system with

high pedestrian traffic. It is difficult to distinguish between

bona fide park ‘visitors’ and ‘non-visitors’. Thus the sample

included both park visitors and pedestrians passing-by the parks.

One out of every three park visitors or passers-by over 16

years old was selected as respondents. Only residents of the six

core districts were invited to participate. Two screening

questions presented at the outset help achieve the age and

residence requirements. We employed sampling quotas, which were

determined by age and gender distribution in the district

concerned (according to the latest available census data), as

these personal characteristics are normally observable from

appearance. Trained interviewers approached a particular age

group and/or gender group when the sampling quota for the

18

remaining age or gender groups had been exhausted. Despite our

every effort to obtain a broadly representative sample, we did

not seek to generalize our findings across the entire population

and the primary objective was to examine the relationship between

behavioral intention and normative factors.

A local university provided logistic support to the survey

and assisted in interviewer recruitment. Five undergraduate

students with prior interview experience were recruited as

interviewers and coached in the procedures of the survey. A

Mandarin-speaking faculty member supervised the fieldwork. Those

individuals who completed the questionnaire received a token as

incentive. Survey activities took place between July and October

2014.

Results

The research team invited 593 Tianjin residents to take part in

the survey in the designated public areas. Response rate was 63%:

375 individuals accepted the request, while 218 refused. There

were 43 withdrawals or incomplete questionnaires, which were

subsequently excluded from analysis. The final sample consists of

332 observations. We processed the data using a standard

statistical package (i.e. STATA) and performed an OLS regression

analysis. The main purpose was to use social capital and risk

perception variables to predict intended economic preparation for

19

urban flood hazards. Further details on the distribution of

responses are available in the Appendix.

Sample characteristics

Respondents’ ages range from 16 to 90, and the average value lies

between 40 and 41. Forty-nine per cent of them are females, and

forty-two per cent hold a university or higher degree. One-third

of respondents (34.5%) have a monthly household income of ¥5,500

– ¥9,000 Chinese yuan (approx. US$891 – US$1,460), and 21.4% have

¥9,000 or above. Less than 15% earn ¥3500 (approx. US$567) or

less per month. There were 20 missing values in monthly income

due to non-response. Ground-floor residents exposed to the risk

of inundation represented 13 per cent of the sample.

Intended economic preparation

Table 1 presents the mean values, standard deviations and factor

loadings of the intended preparation variable. Respondents

indicated varying degrees of economic preparedness for urban

flood hazards. As a widely used tool for risk mitigation,

insurance received the greatest support. Most of the respondents

(71.4 per cent)1 intended to sign up to relevant insurance

schemes to cover their potential losses arising from urban flood

1 Figures presented in this paragraph denote the proportion of respondents selecting the options of ‘Probably’ or ‘Definitely’ in responding to relevant question. See Appendix for details.

20

hazards. More than half (58.6 per cent) would adjust their daily

spending patterns if costs of living vary from normal levels due

to extreme weather. Only one-third (33.7 per cent) would make

contingency arrangements against such events for temporarily

moving their valuable assets to a safer place. Fifty-three per

cent were prepared to identify multiple sources of income. Nearly

half (48.5 per cent) had considered acquiring properties or motor

vehicles in other districts in order to spread the risks of

natural hazards. The five survey items measuring respondents’

economic preparedness were factor analyzed, using the Principal

Component Analysis with Varimax rotation. Results show that these

five variables loaded on one factor (Table 1), suggesting that

they reflect one single latent variable and represent a

particular disposition. Also, these variables collectively

achieved an acceptable Cronbach’s alpha value (0.65), meaning

that the composite scale Intended economic preparation, created by

adding up the five items, is statistically reliable. Higher

values in Intended economic preparation denote greater intention to

prepare for hazards

Hazard experience and risk perception

A small number of the Tianjin residents interviewed (16 per cent)

had recorded some economic losses during a rainstorm or flooding

event in the city. When asked how damaging the last one was to

them and their family, 12.5 per cent of the respondents indicated

21

some damage and about 2 per cent described it as very damaging.

Nevertheless, a lot more people anticipated a certain level of

damage if heavy rainfall and waterlogging at the same magnitude

as in 2012 returned to Tianjin. Twenty seven per cent of

respondents expected some damage, whereas 11 per cent indicated

that impacts on their wellbeing would be significant. On

perceived likelihood of hazards, the majority believed that the

2012 event is likely to return in the next 10 years, i.e. 42.9

per cent selected the ‘maybe’ option and 21.6 per cent indicated

‘possible’ or ‘very likely’. When asked how likely such hazards

would intensify and occur more frequently in the longer term (20

years), 44 per cent responded ‘maybe’ and 15.9 per cent

‘possible’ or ‘very likely’.

Social capital

Most of the individuals expressed some social-normative

considerations (Table 2). More than half believed that they were

expected - by their family (65.6 per cent) and friends (66.2 per

cent) - to prepare for urban flood hazards. Fewer suggested that

such expectation would come from neighbors or other people in

their local residential community (48.9 per cent). As usually

expected, the strength of social ties varies by the type of

relationship. More than 90 per cent had ‘good’ or ‘very good’

relationship with their family. The number went down slightly to

about 87 per cent when friends were concerned. Much fewer (55.4

22

per cent) made the same evaluations for their neighborhood

relationship. The two items on trust revealed a modest level of

division among respondents. While half of them (55.2 per cent)

trusted city planners and administers, nearly one-fifth (19.2 per

cent) expressed skepticism. Local residential committees (jū wěi

huì)2 were seen as less capable of dealing with natural hazards.

Some of the residents we interviewed suggested that these

committees lack substantive power and material resources

essential for organizing effective response to major changes and

catastrophes. Overall, about one-third (31.6 per cent) trusted

their residential committee, but more people (35.9 per cent) held

the opposite view. As shown in Table 2, the social capital

variables were split up into three factors corresponding to their

conceptual categories, which can be called ‘social expectation’,

‘social relationship’ and ‘Institutional Trust’. They managed to

achieve acceptable alpha values and were added up to form three

composite variables accordingly. In general, higher values denote

more social capital.

Relationships between variables

2 In Mainland China, jū wěi huì, which literally means ‘residential committees’, are the most basic form of official community organization. They can be found in every local administrative area in urban China and are managedby government agencies. Their responsibilities include providing residential services, maintaining social order, managing residential environment, and engaging local community in official activities.

23

The regression analysis included all of the survey items

described above. Intended economic preparation was used as the

dependent variable and regressed on respondents’ socio-economic

characteristics, hazard experience and perception, and social

capital items. Descriptive statistics are displayed in Table 3.

Table 4 displays the OLS regression model. The model yielded

an adjusted R2 of 0.39, suggesting that the independent variables

listed in the table explained 39 per cent variations in Intended

economic preparation. Multi-collinearity was not detected as none of

the explanatory variables yielded a Variance Inflation Factor

(VIF) value greater than 5. By usual standards, an explanatory

variable with a p value lower than 0.05 is considered to be

statistically significant.

Only two of the socio-economic variables have significant

impacts on the intention to act, namely, age and income, which

demonstrated negative and positive effects respectively. Younger

and higher-income people had stronger motivation than others to

adopt adaptive measures for coping with severe urban flood

hazards. It is noteworthy that living in a ground-floor unit,

which is a proxy for physical vulnerability, was not a key

factor. Intriguingly, none of the risk-related variables achieved

statistical significance at the 5 per cent level, suggesting that

flood victims and risk-aware individuals were not more likely to

make advance arrangements for future hazards. On the other hand,

all of the three social capital variables indicated moderately

24

strong and positive effects on intention. Those individuals who

were subject to greater social expectations, had better

relationships with other social actors, and tend to trust

relevant institutions were more likely to adopt resilient

economic measures.

Discussion and conclusions

Social capital is crucial for enhancing community and household

preparedness for shocks and abrupt environmental changes.

Productive and trusted relations among social and organizational

actors and the effectual social norms engendered are a vital

source of economic resilience to environmental change. But social

capital is not a universal common good and may create perverse

incentives and a false sense of security, leading to delayed

action or maladaptation (Pelling, 2011; Wolf et al., 2010).

Inquiry into its complex implications for hazard adjustment and

adaptation remains an incomplete project.

The pivotal role of social capital warrants greater efforts

by Chinese researchers. Existing local studies tend to put more

emphasis on human cognition and hazard experience than social-

normative and relational dimensions. For instance, Huang et al.

(2010), Ge et al. (2011) and Yu et al. (2013) conclude that what

the Chinese government needs to do is to provide more hazard

information to the general public and improve knowledge-based

communication of risk. Zheng and Dallimer (2015) have made

25

similar recommendations, although they go further by bringing up

personal-subjective factors. Many Chinese studies on public

response follow the ‘information-deficit’ model (Lo, et al.,

2012; Owens, 2000). The idea that low perception and awareness of

risk is the main barrier to enhancing community preparedness for

disruptive environmental change is well received.

Our study contributes to these debates by investigating the

systematic relationship between social capital and household

economic resilience in the context of a flood-prone port city in

China. It demonstrated the statistical effects of social capital

and risk perception on the intention of Tianjin residents to

adopt resilient economic measures for coping with urban flood

hazards. We found no observable evidence supporting the standard

assumption that perceived risk and damage experience positively

contribute to hazard adjustments. Consistent with these results,

we noted that ground-floor residents, who are particularly

vulnerable to waterlogging and inundation, did not indicate

stronger motivation to act than others. Our first research

hypothesis is therefore not supported. Nonetheless these findings

corroborate those of Bubeck et al. (2012, 2013) and Lo (2013),

and imply that merely providing more hazard information and

improving public understanding of risk are insufficient.

Alternative strategies are needed to make management

interventions effective.

26

Utilizing social capital is a promising way for engaging

households in hazard risk mitigation and climate change

adaptation (Frank et al., 2011; Bihari and Ryan, 2012; Lo, 2013).

Wang, et al. (2012) have suggested that risk managers in China

should give higher priority to strengthening residents’ social

networks and their engagement with local community as a means to

improve community preparedness for natural hazards. Our study

provides empirical support to this recommendation by

demonstrating that social capital has significant positive

impacts on economic preparation for catastrophic rainstorm and

flooding. Residents’ intention to make advance arrangements for

reducing potential economic losses due to such hazards increased

with the levels of social expectation, social relationship, and

organization trust. The more social capital one has accumulated,

the stronger motivation to act. This offers evidence for our

second hypothesis and indicates the importance of building closer

ties and trusted relations among social and institutional actors

for enhancing household and community resilience. Strengthening

these aspects of people’s social life will be conducive to

reducing their economic losses and ultimately the costs borne by

government agencies in the prospects of more intense and frequent

natural hazards in the coming decades.

The findings echo the calls by international researchers for

addressing the link between climate change adaptation and

prevailing social structures and institutions. Social networks

and trusted relations facilitate the exchange of information and

27

resources, and provide greater access to opportunities that can

help absorb economic shocks and reduce risks. These networks and

associated flows of information and resources between individuals

and groups oil the wheels of decision making (Adger, 2000, 2003).

It is imperative to build capacity for social collectives and

local community groups to respond to climate change (Pelling and

High, 2005; Pelling, 2011). Based on the results of the Tianjin

survey, we suggest that the accumulation of social capital

through trusted networks and institutions can help enhance

Chinese residents’ preparedness for the consequences of

accelerating climate change. This requires greater efforts on

strengthening the adaptation functions of formal and informal

communal institutions, such as residential committees and mutual-

aid cooperatives, respectively. Although these social

considerations and measures have proven elsewhere to be important

for enhancing resilience, the unique context of China has created

a multitude of hurdles.

Our recommendation points to a governance issue that

requires structural reforms. An enduring problem in the present-

day China is that social cohesion among urban residents has

declined over the past few decades, as a consequence of

increasing internal migration, income disparities, and uneven

development. Our interviews with local residents reveal that they

had little trust in people outside their immediate social

circles, and this appears to be a matter of social ethos rather

than isolated individual dispositions. Moreover, the Chinese

28

government continues to impose many restrictions on non-state

social organizations, posing constraints on the organic growth of

the civil society. Consequently, the number and capacity of

informal social institutions within cities are limited. As a

frontline government agency, residential committees could play a

more active and important role in hazard risk management, but

they lack authority and resources to assume more responsibilities

and augment their capacity in social organization. The

conventional top-down approach of social governance in China

constitutes a hurdle for strengthening the social function of

local organizations that are more capable of fostering social

connections and enhancing household adaptation.

Apart from moving beyond the information-deficit model,

decision-makers in China should allow greater empowerment in

social participation and build adaptive capacity at the

‘grassroots’ level. Our findings therefore lend support to the

bottom-up approaches of governing flood risks. Community-based

social innovations, such as encouraging public-public

partnerships (e.g. between non-governmental cooperatives and

government agencies) and engaging long-term residents in

volunteer stewardship programs (Bihari and Ryan, 2012; Suckall et

al., 2014), are much welcomed.

On methodology, we suggest that further research based on

social surveys should include objective, behavioral indicators of

social capital, such as the individuals’ membership in social

29

organizations and trade unions, the number of neighbors they

know, the frequencies of participating in church activities and

political elections, etc. Individual household surveys can hardly

capture the notion of social capital in its entirety, which is a

property of the society at large and embedded into social

structures and institutions. Survey designers need to recognize

the multiple dimensions of the concept and the various sites for

socialization in which it accumulates. Although extensive surveys

have their limits in fully capturing the concept, behavioral

indicators could at least mitigate the inherent ambiguities and

inconsistencies in self-assessing the quality of social

relationship and the level of trust. The use of alternative

measures would be useful for cross-checking the results and could

strengthen the argument for addressing social dimensions.

Acknowledgements

This research was funded by the Academy of the Social Sciences in

Australia under the Australia-China Joint-action Program and the

Griffith Climate Change Response Program at Griffith University.

The authors appreciate the research assistance provided by

students at the Tianjin Normal University.

30

References

Adger, W. N. (2000). Institutional Adaptation to Environmental Riskunder the Transition in Vietnam. Annals of the Association of AmericanGeographers, 90 (4), 738-758.

Adger, W. N. (2003). Social Capital, Collective Action, and Adaptationto Climate Change. Economic Geography, 79 (4), 387-404.

Alinovi, L., Mane, E., & Romano, D. (2009). Measuring householdresilience to food insecurity: application to Palestinianhouseholds. In EC-FAO Food Security Programme: Linking Information andDecision Making to Improve Food.

Barnett, A. J., & Eakin, H. C. (forthcoming). “We and us, not I andme”: Justice, social capital, and household vulnerability in aNova Scotia fishery. Applied Geography.

Bihari, M., & Ryan, R. (2012). Influence of social capital oncommunity preparedness for wildfires. Landscape and Urban Planning, 106(3), 253-261.

Bubeck, P., Botzen, W. J. W., & Aerts, J. C. J. H. (2012). A Review ofRisk Perceptions and Other Factors that Influence FloodMitigation Behavior. Risk Analysis, 32 (9), 1481-1495.

Bubeck, P., Botzen, W. J. W., Kreibich, H., & Aerts, J. C. J. H.(2013). Detailed insights into the influence of flood-copingappraisals on mitigation behaviour. Global Environmental Change, 23(5), 1327-1338.

Chan, F., Mitchell, G., & McDonald, A. T. (2012). Flood risk appraisaland management in mega-cities: a case study of practice in thePearl River Delta, China. Water Practice & Technology, 7 (4).

Coleman, J. (1990). Foundations of social theory. Harvard University Press,Boston, MA.

Eriksen, S., & Selboe, E. (2012). The social organisation ofadaptation to climate variability and global change: The case ofa mountain farming community in Norway. Applied Geography, 33 (0),159-167.

Frank, E., Eakin, H., & López-Carr, D. (2011). Social identity,perception and motivation in adaptation to climate risk in thecoffee sector of Chiapas, Mexico. Global Environmental Change, 21 (1),66-76.

Fuchs, R., Conran, M., & Louis, E. (2011). Climate Change and Asia’sCoastal Urban Cities. Environment and Urbanization Asia, 2 (1), 13-28.

Ge, Y., Xu, W., Gu, Z.-H., Zhang, Y.-C., & Chen, L. (2011). Riskperception and hazard mitigation in the Yangtze River Deltaregion, China. Natural Hazards, 56 (3), 633-648.

31

Goulden, M. C., Adger, W. N., Allison, E. H., & Conway, D. (2013).Limits to Resilience from Livelihood Diversification and SocialCapital in Lake Social–Ecological Systems. Annals of the Association ofAmerican Geographers, 103 (4), 906-924.

Güneralp, B., Güneralp, İ., & Liu, Y. (2015). Changing global patternsof urban exposure to flood and drought hazards. Global EnvironmentalChange, 31 (0), 217-225.

Han, S.-q., Xie, Y.-y., Li, D.-m., Li, P.-y., & Sun, M.-l. (2006).Risk analysis and management of urban rainstorm water logging inTianjin*. Journal of Hydrodynamics, Ser. B, 18 (5), 552-558.

Huang, L., Duan, B., Bi, J., Yuan, Z., & Ban, J. (2010). Analysis ofDetermining Factors of the Public's Risk Acceptance Level inChina. Human and Ecological Risk Assessment: An International Journal, 16 (2),365-379.

Keil, A., Zeller, M., Wida, A., Sanim, B., & Birner, R. (2008). Whatdetermines farmers’ resilience towards ENSO-related drought? Anempirical assessment in Central Sulawesi, Indonesia. ClimaticChange, 86 (3-4), 291-307.

Kunreuther, H. (1996). Mitigating disaster losses through insurance.Journal of Risk and Uncertainty, 12 (2), 171-187.

Kunreuther, H. (2006). Disaster Mitigation and Insurance: Learningfrom Katrina. Annals of the American Academy of Political and Social Science, 604,208-227.

Le Dang, H., Li, E., Nuberg, I., & Bruwer, J. (2014). Farmers’assessments of private adaptive measures to climate change andinfluential factors: a study in the Mekong Delta, Vietnam. NaturalHazards, 71 (1), 385-401.

Lo, A., & Cheung, L. O. (2015). Seismic risk perception in theaftermath of Wenchuan earthquakes in southwestern China. NaturalHazards, 78 (3), 1979-1996.

Lo, A. Y. (2013). The Role of Social Norms in Climate Adaptation:Mediating Risk Perception and Flood Insurance Purchase. GlobalEnvironmental Change, 23 (5), 1249–1257.

Lo, A. Y., Chow, A. T., & Cheung, S. M. (2012). Significance ofPerceived Social Expectation and Implications to ConservationEducation: Turtle Conservation as a Case Study. EnvironmentalManagement, 50 (5), 900-913.

Miceli, R., Sotgiu, I., & Settanni, M. (2008). Disaster preparednessand perception of flood risk: A study in an alpine valley inItaly. Journal of Environmental Psychology, 28 (2), 164-173.

National Bureau of Statistics of China. (2014). National Database. In.National Bureau of Statistics of China, Bejing

32

Norgaard, K. M. (2011). Living in Denial: Climate Change, Emotions, and Everyday Life.MIT Press Cambridge, Mass.

Owens, S. (2000). 'Engaging the public': Information and deliberationin environmental policy. Environment and Planning A, 32, 1141-1148.

Pelling, M. (2011). Adaptation to climate change : from resilience to transformationRoutledge, Oxon, U.K.

Pelling, M., & High, C. (2005). Understanding adaptation: What cansocial capital offer assessments of adaptive capacity? GlobalEnvironmental Change, 15, 308-319.

Pretty, J., & Ward, H. (2001). Social Capital and the Environment.World Development, 29 (2), 209-227.

Putnam, R. D. (1995). Bowling Alone: America's Declining SocialCapital. Journal of Democracy, 6 (1), 65-78.

Rose, A. (2004). Defining and measuring economic resilience todisasters. Disaster Prevention and Management, 13 (4), 307-314.

Rose, A. (2007). Economic resilience to natural and man-madedisasters: Multidisciplinary origins and contextual dimensions.Environmental Hazards, 7 (4), 383-398.

Suckall, N., Tompkins, E., & Stringer, L. (2014). Identifying trade-offs between adaptation, mitigation and development in communityresponses to climate and socio-economic stresses: Evidence fromZanzibar, Tanzania. Applied Geography, 46 (0), 111-121.

Tianjin Bureau of Statistics. (2014). 2013 National Economic andSocial Development in Tianjin. In. Tianjin Bureau of Statistics,Tianjin.

Vari, A., Linnerooth-Bayer, J., & Ferencz, Z. (2003). StakeholderViews on Flood Risk Management in Hungary's Upper Tisza Basin.Risk Analysis, 23 (3), 585-600.

Varis, O., Kummu, M., & Salmivaara, A. (2012). Ten major rivers inmonsoon Asia-Pacific: An assessment of vulnerability. AppliedGeography, 32 (2), 441-454.

Wang, G., Liu, Y., Wang, H., & Wang, X. (2014). A comprehensive riskanalysis of coastal zones in China. Estuarine, Coastal and Shelf Science,140 (0), 22-31.

Wang, M.-Z., Amati, M., & Thomalla, F. (2012). Understanding thevulnerability of migrants in Shanghai to typhoons. Natural Hazards,60 (3), 1189-1210.

Wilson, G. A. (2015). Community Resilience and Social Memory.Environmental Values, 24 (2), 227-257.

Wolf, J., Adger, W. N., Lorenzoni, I., Abrahamson, V., & Raine, R.(2010). Social capital, individual responses to heat waves andclimate change adaptation: An empirical study of two UK cities.Global Environmental Change, 20 (1), 44-52.

33

World Bank. (2010). Natural Hazards, UnNatural Disasters: TheEconomics of Effective Prevention. In. World Bank, WashingtonD.C.

Wu, Q. Z. (2012). Discussion on rains and floods in Beijing and urbanwaterlogging prevention. China Ancient City, 10, 4-13 [in Chinese].

Xie, X. L., Lo, A. Y., Zheng, Y., Pan, J., & Luo, J. (2014). Genericsecurity concern influencing individual response to naturalhazards: evidence from Shanghai, China. Area, 46 (2), 194-202.

Yu, H., Wang, B., Zhang, Y.-J., Wang, S., & Wei, Y.-M. (2013). Publicperception of climate change in China: results from thequestionnaire survey. Natural Hazards, 69 (1), 459-472.

Zheng, Y., & Dallimer, M. (2015). What motivates rural households toadapt to climate change? Climate and Development, 1-12.

34

Table 1 Factor analysis of economic preparation variables

Mean S.D. Factorloading

Cronbach’s alpha

F1. Take out insurance 3.37 1.23 0.534

0.654

F2. Reduce expenses 3.81 1.13 0.716F3. Arrange storage for valuables

2.85 1.15 0.615

F4. Diversify sources of income

3.16 1.24 0.709

F5. Spread assets across different places

3.10 1.26 0.655

Eigenvalues 2.106% of variance explained 42.121Higher mean values denote stronger intention to act

Factor extraction method: Principal Component Analysis with Varimax rotation.

35

Table 2 Factor analysis of social capital variables

Mean S.D.

Factor loading Cronbach’salphaSocial

expectation

Socialrelation

ship

InstitutionalTrust

A1. Family expectation 3.64 0.99

.780

0.693A2. Friends expectation 3.65 0.88

.837

A3. Community expectation 3.29 0.95

.669

B1. Relationship with family 4.49 0.66

.835

0.619B2. Relationship with friends 4.21 0.70

.832

B3. Relationship with community members

3.64 0.74

.535

C1. Trust in municipal government

3.47 1.03

.836

0.676C2. Trust in local residential committee

2.95 1.07

.836

Eigenvalues 1.857 1.749 1.574% of variance explained 23.209 21.858 19.675Higher mean values denote greater expectation (A1-A3), better relationship

(B1-B2) and more trust (C1-C2)

Factor extraction method: Principal Component Analysis with Varimax rotation.

36

Table 3 Descriptive statistics of model variables

VariableDescription Range Mean S.D

.Socio-economic characteristicsAge Open-ended 16 –

9040.5

816.78

University degree

1 = University degree or higher, 0 =otherwise

0 - 1 0.42 0.49

Female 1 = Female, 0 = Male 0 - 1 0.49 0.50

Monthly household income

1 = $2,500 or lower, 2 = $2,500 – $3,500 …… 6 = $9,000 or higher (in Chinese yuan)

1 - 6 5.28 1.53

Ground-floor resident

1 = Current home located at the ground floor of a residential building, 0 = Otherwise

0 - 1 0.13 0.34

Hazard experience and perceptionDamage experience

1 = had suffered economic damage dueto rainstorms or floods, 0 = Otherwise

0 - 1 0.16 0.36

Level of damage (actual)

1 = No damage, 2 = Minor damage …… 5= Extremely damaging

1 - 5 1.44 0.81

Level of damage (expected)

1 = No damage, 2 = Minor damage …… 5= Extremely damaging

1 - 5 2.30 0.98

Perceived likelihood (recurrence)

1 = Very unlikely, 2 = Not quite possible …… 5 = Very likely

1 - 5 2.85 0.99

Perceived likelihood (intensification)

1 = Very unlikely, 2 = Not quite possible …… 5 = Very likely

1 - 5 2.76 0.86

Social capitalSocial expectation

Composite scale (A1 – A3). Higher values denote greater expectations

3 -15

10.58

2.21

Social relationship

Composite scale (B1 – B3). Higher values denote better relationships

3 -15

12.34

1.58

Institutional Trust

Composite scale (C1 – C2). Higher values denote higher level of trust

2 –10

6.42 1.83

Dependent variableIntended Composite scale (F1 – F5). Higher 5 – 16.3 3.8

37

economic preparation

values denote greater intention to prepare for hazards

24 4 7

38

Table 4 OLS Regression Analysis

Coefficient

Standard Error

95% ConfidenceInterval

Age -0.088 ** 0.012 -0.113-

0.064University degree 0.076 0.417 -0.745 0.897Female -0.572 0.356 -1.273 0.130Monthly household income 0.446 ** 0.125 0.200 0.693Ground-floor resident -1.016 0.536 -2.071 0.039Damage experience -0.172 0.671 -1.492 1.149Level of damage (actual) 0.226 0.319 -0.401 0.853Level of damage (expected) 0.330 0.198 -0.059 0.720Perceived likelihood (recurrence) 0.405 0.219 -0.025 0.836Perceived likelihood (intensification) 0.334 0.239 -0.136 0.803Social expectation 0.214 * 0.100 0.016 0.412Social relationship 0.274 * 0.123 0.031 0.516Institutional Trust 0.239 * 0.106 0.030 0.448

(Constant) 8.490 1.843 4.86212.11

9Adj. R2 0.394F (13, 274) 15.330Prob > F 0.000Number of observations 288Dependent variable: Intended economic preparation

** and * denote significance at .01 and .05 levels respectively.

39

Appendix

Table A1 Distribution of respondents by intended economic preparation

Intended economic preparation Percentage of respondents (%)Definitelynot (1) /

probably not(2)

Notsure(3)

Probably(4) /

definitely (5)

F1. If you could, how likely would you take out insurance to cover yourpotential losses arising from urban flood hazards?

15.5 13.1 71.4

F2. If catastrophic rainstorms or flooding increase your daily expenses, how likely would you try to reduce your expenses without compromising standard of living?

28.6 12.8 58.6

F3. If you could, how likely would you arrange temporary storage in advance for your valuables when a major catastrophe event occurs?

46.5 19.8 33.7

F4. How likely would you seek out multiple sources of income?

33.7 13.4 52.9

F5. If you could, how likely would you acquire properties or motor vehicles in other Districts in orderto spread the financial risks of natural hazards?

36.1 15.5 48.5

Measured on a five-point Likert scale

40

Table A2 Distribution of respondents by level of damage due to catastrophic rainstorms or flooding

Level of damage Percentage of respondents (%)No damage

(1) / minordamage (2)

Somedamage(3)

Fairlydamaging (4)/ extremelydamaging (5)

Actual: How damaging were the last major catastrophic rainstorms or flooding to you and your family?

85.6 12.5 1.9

Expected: How much personal damage would a major catastrophic event like the 2012 rainstorms cause?

61.6 27.4 10.9

Measured on a five-point Likert scale

41

Table A3 Distribution of respondents by perceived likelihood of catastrophic rainstorms or flooding

Perceived likelihood of urban flood hazards

Percentage of respondents (%)

Very likely(1) / not

quitepossible (2)

Maybe(3)

Possible (4)/ very

likely (5)

Recurrence: How likely is Tianjinto experience a major catastrophic event like the 2012 rainstorms in the next 10 years?

35.6 42.9 21.6

Intensification: Do you think these rainstorms or flooding would become more frequent and severe in the next 20 years?

40.1 44.0 15.9

Measured on a five-point Likert scale

42

Table A4 Distribution of respondents by social expectation

Social expectation Percentage of respondents (%)Definitelynot (1) /

probably not(2)

Not sure(3)

Probably(4) /

definitely (5)

A1. Do you think your family would encourage you to prepare for urban flood hazards?

15.5 18.8 65.6

A2. Do you think your friends would encourage you to prepare for urban flood hazards?

12.2 21.6 66.2

A3. Do you think other people in your residential community would encourage you to prepare for urban flood hazards?

22.0 29.1 48.9

Measured on a five-point Likert scale

43

Social relationship Percentage of respondents (%)Very bad

(1)Bad(2)

Fair(3)

Good(4)

Very good(5)

B1. How is your relationship with your family?

0.6 0.0 5.5 37.7 56.2

B2. How is your relationship with your friends?

0.3 0.6 12.2 51.4 35.6

B3. How is your relationship with other people in your residential community ?

1.2 0.6 42.9 43.5 11.9

Table A5 Distribution of respondents by social relationship

Measured on a five-point Likert scale

44

Table A6 Distribution of respondents by Institutional Trust

Institutional Trust Percentage of respondents (%)Absolutelynot (1) /not likely

(2)

Not sure(3)

Likely(4) /

Absolutely (5)

C1. Do you trust that the municipal government can help residents cope with the impacts of urban flood hazards?

19.2 25.6 55.2

C2. Do you trust that your local residential committee can help residents cope with the impacts of urban flood hazards?

35.9 32.5 31.6

Measured on a five-point Likert scale

45

Caption for figures

Figure 1: Map of Tianjin

Figure 2: View from a medium-rise building over the Yuexiu Road Community in the Hexi District of Tianjin, a megacity in China. In the old core districts, buildings are closely packed and humanactivities are highly concentrated. Both residents and infrastructures are vulnerable to waterlogging and inundation in the event of flooding and/or rainstorm.

Figure 3: View from a typical alley in Dongshezhai Community in the Hexi District of Tianjin, which is located in close proximity to a major river called ‘Haihe’. Residents in the old core districtsoften park their cars and bikes along curbside, with minimal protection from floods, due to limited space within the private residential area. Lower-floor residents are particularly vulnerable due to their location and the concentration of assets.

46